Differences

This shows you the differences between two versions of the page.

--- aira:start [2025/11/03 14:29] – [Schedule Autumn 2025] mzk
+++ aira:start [2025/11/03 14:36] (current) – [2025-10-30] mzk
@@ Line 471: / Line 471: @@
+==== 2025-11-06 ====
+<WRAP column 15%>
+{{ :aira:kliegr-sykora-foto.jpg?200| }}
+</WRAP>
+<WRAP column 75%>
+**Speaker**: Tomáš Kliegr and Lukas Sykora @ Prague University of Economics and Business
+**Title**: LLM-based feature generation from text for interpretable machine learning.
+**Abstract**:
+Traditional text representations like embeddings and bag-of-words hinder rule learning and other interpretable machine learning methods due to high dimensionality and poor comprehensibility. This article investigates using Large Language Models (LLMs) to extract a small number of interpretable text features. We propose two workflows: one fully automated by the LLM (feature proposal and value calculation), and another where users define features and the LLM calculates values. This LLM-based feature extraction enables interpretable rule learning, overcoming issues like spurious interpretability seen with bag-of-words. We evaluated the proposed methods on five diverse datasets (including scientometrics, banking, hate speech, and food hazard). LLM-generated features yielded predictive performance similar to the SciBERT embedding model but used far fewer, interpretable features. Most generated features were considered relevant for the corresponding prediction tasks by human users. We illustrate practical utility on a case study focused on mining recommendation action rules for the improvement of research article quality and citation impact.
+**Biogram**:
+Tomáš Kliegr is a Professor at the Faculty of Informatics and Statistics at the Prague University of Economics and Business (VSE Praha), where he is part of the Data Science & Explainable AI (DSXAI) research team. His research interests include Explainable AI (XAI), Interpretable Machine Learning, and neurosymbolic methods. He has published on topics such as the effect of cognitive biases on model interpretation in journals including Artificial Intelligence and Machine Learning. He is active in the rule-based systems community.
+Dr. Lukas Sykora is a Research Assistant at the Department of Information and Knowledge Engineering and a Lecturer at the Prague University of Economics and Business.  He holds a PhD in Applied Informatics (2025), where his doctoral thesis focused on action rule mining. He has authored several publications on this topic, including "Apriori Modified for Action Rules Mining." He also brings industry experience as a Solution Architect Team Lead at Ogilvy.
+</WRAP>
+<WRAP clear></WRAP>
 ==== 2025-10-30 ====